A performance evaluation of local descriptors

By Krystian Mikolajczyk, Cordelia Schmid. 2005

8,513 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

Aggregating Local Image Descriptors into Compact Codes (2012, methodology): “First, this representation benefits from powerful local descriptors, such as the SIFT descriptor [7] or more recent ones [8], [9], [10].”
Understanding Image Representations by Measuring Their Equivariance and Equivalence (2014, neutral): “One of the common approaches is to sample the nuisance feature space sparsely with various “detectors”—such as local feature detectors with different normalization schemes (Lindeberg 1998; Lowe 1999; Mikolajczyk and Schmid 2003), bounding box proposals (Uijlings et al.”
Learning to compare image patches via convolutional neural networks (2015, methodology): “Evaluation on the Mikolajczyk dataset [16] showing the mean average precision (mAP) averaged over all types of transformations in the dataset (as usual, the mAP score measures the area under the precision-recall curve).”
Generic 3D Representation via Pose Estimation and Matching (2016, background): “scribing, and matching image features, either through a handcrafting the feature [35,11,39,38,5,40] or learning it [50,25,65,9,66,49,19].”
Strong Appearance and Expressive Spatial Models for Human Pose Estimation (2013, background): “Various local appearance models have been proposed, including stretchable models representing local appearance of body joints [34, 31, 26] and cardboard models modelling appearance of body parts as rigid templates [2, 25, 16].”
Fracking Deep Convolutional Image Descriptors (2014, background): “Normalization has been shown to be important for deep networks (Jarrett et al., 2009) and fundamental for descriptors (Mikolajczyk & Schmid, 2005).”
Random Walks for Image Segmentation (2006, background): “Assign local descriptors [25] to each pixel independently.”
Pictorial structures revisited: People detection and articulated pose estimation (2009, methodology): “Our discriminative part detectors densely sample a variant of the shape context descriptor initially proposed in [14] and previously used for pedestrian detection [19].”
Detection and Localization of Multiple Objects (2006, neutral): “descriptors in a comparative study [5].”
Local Features and Kernels for Classification of Texture and Object Categories: A Comprehensive Study (2006, positive): “The SIFT descriptor (Lowe, 2004) has been shown to outperform a set of existing descriptors ( Mikolajczyk and Schmid, 2005 ), while SPIN and RIFT, introduced by Lazebnik et al. (2005), have achieved good performance in the context of texture classification.”
SAR-SIFT: A SIFT-Like Algorithm for SAR Images (2015, methodology): “Gradient location and orientation histogram [3] is computed on a log-polar grid and upon 17 sectors, the”
The Generalized PatchMatch Correspondence Algorithm (2010, methodology): “The common way to deal with this case is via keypoint detectors [4].”
Learning hierarchical invariant spatio-temporal features for action recognition with independent subspace analysis (2011, methodology): “Typical examples of such successful features for static images are SIFT [24] [25], HOG [4], GLOH [27] and SURF [1].”
Recognizing Objects in 3D Point Clouds with Multi-Scale Local Features (2014, methodology): “The NNDR-based matching technique outperforms the other two techniques in terms of matching precision due to the fact that it penalizes the descriptors, which have many similar matches [23].”
ASIFT: A New Framework for Fully Affine Invariant Image Comparison (2009, methodology): “A two-resolution scheme will further reduce the ASIFT complexity to about twice that of SIFT.”
SuperPoint: Self-Supervised Interest Point Detection and Description (2017, background): “Traditional interest point detectors have been thoroughly evaluated [24, 16].”
A Comprehensive Performance Evaluation of 3D Local Feature Descriptors (2015, methodology): “When compared to the performance evaluations of 2D keypoint detectors (Schmid et al. 2000; Mikolajczyk and Schmid 2004; Mikolajczyk et al. 2005; Moreels and Perona 2007), 2D local feature descriptors (Mikolajczyk and Schmid 2005; Moreels and Perona 2005, 2007; Burghouts and Geusebroek 2009), and 3D…”